From Code Completion to Autonomous Agents: How Python Teams Can Transform Their Engineering Practices

The rapid evolution of AI-assisted development has fundamentally changed how engineering teams approach software development. Jon Hyman, co-founder and CTO of Braze, recently shared on Stack Overflow's Leaders of Code podcast how his organization transformed into an AI-first team in just a few months, with remarkable results that offer valuable lessons for Python development teams.

From Mobile-First to AI-First: A 15-Year Journey

Hyman has led Braze's engineering organization for nearly 15 years, guiding the company from startup to global leader in customer engagement platforms. His leadership approach has evolved significantly with the company's growth:

"When we founded the company almost 15 years ago, that is basically a lifetime in technology," Hyman explained. "It's been really exciting to be from the birth of mobile and really the growth of that and build a business that was catalyzed by the changes to the world for mobile and now be at the time of being in a world that's being catalyzed and changed by AI."

Initially, Braze operated with a small team where everyone worked on everything. As the company grew, they evolved to specialized teams and eventually implemented a divisional structure with engineering managers, directors, VPs, and other leadership roles. This structure has proven essential for managing the complexity of a 300-person engineering organization.

The AI Transformation: From Code Completion to Autonomous Agents

Braze's AI journey began with familiar tools like GitHub Copilot for code completion. However, the real transformation began in August 2025 when the company decided to build an MCP (Model Context Protocol) server entirely with AI.

"We actually said, 'You know what? The MCP server is something that I'd say is like low stakes. It's a greenfield. Let's just try and build the MCP server entirely with AI. Let's take two engineers and have them do it,'" Hyman recalled. "What they did is they built the MCP server route, they ended up about six weeks ahead of schedule because of using AI. And it was really exciting because that for me was the moment where I actually got that click of, oh gosh, this is going to be a stepwise increase in efficacy for our teams."

The breakthrough came with the release of Opus 4.5 in November 2025, which Hyman describes as "the first model that I worked with that felt like it needed little direction or correction in order to build with a great meaningful feature quickly and correctly."

The results have been dramatic: "Last week, I was looking at the stats. Over 60% of the code we committed to our main repositories was written with AI."

Practical Strategies for Python Development Teams

For Python development teams looking to adopt similar approaches, several key strategies emerge from Braze's experience:

1. Start with Enablement and Experimentation

Braze's journey began with a focus on enabling the organization through access to tools and education:

"Over the next two months from February to call it late April, early May, I was trying to look at how I can just give the team access to different coding tools. We bought Cursor licenses for folks. We doubled down on GitHub Copilot for people who wanted it and started doing automated code reviews. I got Claude Code access for anyone who wanted to use that through AWS Bedrock. And just we started doing AI lunch and learns in order to enable the team."

For Python teams, this could involve:

• Setting up GitHub Copilot with Python-specific configurations
• Providing access to tools like Cursor or CodeWhisperer optimized for Python
• Creating internal documentation on Python best practices that AI tools can reference
• Organizing workshops on using AI for common Python tasks like data processing with pandas or web development with Django/Flask

2. Focus on High-Impact, Low-Stakes Projects

The MCP server project was successful because it was "low stakes" and "greenfield." Python teams can identify similar opportunities:

"We tried to do the general things you do in engineering. Have some lunch and learns, try to do some show and tells, say that there's people available to help you get set up and running and promote the things that are working really well and try to get that bottoms up groundswell that I think is really effective in engineering because engineers like to tinker, they like to hear what other people are doing."

For Python teams, this might involve:

• Building internal tools for data analysis using Python and AI
• Creating automated documentation generators for Python APIs
• Developing testing frameworks that use AI to generate test cases
• Creating Python packages with AI-assisted development

3. Python-Specific AI Workflows

Python's rich ecosystem offers unique opportunities for AI-assisted development:

• Data Science: AI tools can help with data cleaning, feature engineering, and model selection in Python projects using pandas, NumPy, and scikit-learn
• Web Development: Frameworks like Django and Flask can benefit from AI-assisted route creation, database schema generation, and API design
• DevOps: Python scripts for infrastructure management can be generated and refined with AI assistance
• Testing: AI can help generate pytest test cases, mock objects, and edge case scenarios

Hyman emphasized the importance of codifying Python-specific best practices: "We're trying to write that down because we're seeing things like the scaffolding of endpoint creation in the front end is just abysmal because it's kind of this, what I'd say, like vanilla off the shelf stuff that doesn't fit into how we've built the application."

For Python teams, this means creating detailed documentation on:

• Project structure conventions
• Code style guidelines (following PEP 8)
• Testing methodologies
• Dependency management practices
• Security considerations

Managing Challenges and Costs

One significant challenge Braze faced was the cost of AI inference:

"One of the big challenges that we're running into in engineering is we're realizing just how expensive AI inference is. Ultimately, these models are very good now, but they cost a lot of money, especially when people are using them every day as part of their jobs."

Hyman noted that one engineer had already spent $150 on inference in a single day. For Python teams, this means:

1. Optimize AI Usage for Python Workflows

Not all Python tasks require the same level of AI capability. Teams can:

• Use lighter models for simple code completion tasks
• Reserve more powerful models for complex architectural decisions
• Implement caching for common Python patterns and solutions
• Use specialized AI tools for specific Python domains (e.g., data science, web development)

2. Measure Productivity and Value

Hyman outlined three streams for measuring AI success:

1. Enablement and awareness: Getting teams familiar with tools
2. AI as coworker: Moving from assistance to autonomous task completion
3. Changing business metrics: Impact on top and bottom line

For Python teams, this could involve:

• Tracking time saved on common Python tasks
• Measuring code quality metrics (test coverage, bug rates)
• Monitoring developer satisfaction and productivity
• Assessing business impact of AI-accelerated Python development

3. Balance Between Speed and Quality

While AI can accelerate Python development, maintaining code quality remains crucial:

"You can't vibe code scale. Again, vibe scale, really. And the things is that you can build a lot of, get a lot of code out there, but ultimately being able to run that at high complexity, high scale, high downy use cases is something that requires deep understanding of what you're doing, the business problem that you're solving, and then how all the systems work together."

For Python teams, this means:

• Implementing rigorous review processes for AI-generated code
• Maintaining comprehensive test suites
• Ensuring adherence to Python best practices and architectural patterns
• Preserving institutional knowledge about the system

The Future of AI-Assisted Python Development

Looking ahead, Hyman envisions a future where AI agents work continuously:

"I think that the world's going to go ahead where all of our teams are going to have agents that are building stuff 24/7, that are responding to bug reports, that are responding to product questions, that are helping build roadmap, people are going to go to sleep and wake up with features being built by AI."

For Python development, this future includes:

1. Specialized AI Agents for Python Ecosystem

We're likely to see AI agents specifically trained on Python codebases, frameworks, and best practices. These agents could:

• Understand Python-specific conventions and idioms
• Generate code optimized for specific Python libraries
• Provide context-aware suggestions based on the Python ecosystem
• Automate Python-specific tasks like dependency management and virtual environment setup

2. Enhanced Python Development Environments

Tools like VS Code and PyCharm will likely integrate more sophisticated AI assistance specifically for Python:

• Intelligent code completion with deep Python context
• Automated refactoring suggestions following Python best practices
• Real-time code quality checks specific to Python conventions
• Seamless integration with Python testing frameworks

3. Knowledge Management for Python Projects

As AI becomes more involved in development, managing institutional knowledge becomes even more critical:

"I think that this is leading to this bunch of spaghetti skills and ways of working, and we're going to have this challenge on our hands that I think we already have, but haven't really said out loud where people are using different tools, they have different sets of skills that are in each of those tools."

For Python teams, this means:

• Creating comprehensive documentation of Python codebases
• Developing standardized patterns for common Python tasks
• Implementing knowledge sharing mechanisms
• Building internal Python expertise to guide AI-assisted development

Practical Implementation for Python Teams

Based on Braze's experience, here's a practical roadmap for Python teams looking to adopt AI-assisted development:

1. Start small: Identify a specific Python project or task where AI assistance could provide quick wins
2. Invest in tooling: Set up AI development tools optimized for Python (GitHub Copilot, Cursor, etc.)
3. Create guidelines: Develop Python-specific standards for AI-assisted development
4. Measure impact: Track metrics like code quality, development speed, and developer satisfaction
5. Scale gradually: Expand AI assistance to more complex Python projects as confidence grows
6. Plan for costs: Implement strategies to manage AI inference expenses
7. Focus on quality: Maintain rigorous testing and review processes for AI-generated Python code

Hyman's advice for engineering leaders is clear: "If you're already not launching a bunch of agents in parallel and having them work on different features, I would say go do something like that. If you don't have your AI app on your home screen or your phone, go ahead and put it there. So then that way your first instinct is to open that up instead of going to Google and doing a search, really just start engaging with the AI."

For Python development teams, the message is similar: Start experimenting with AI tools in your Python workflows today, learn from the experience, and be prepared to evolve your practices as AI capabilities continue to improve. The future of Python development is not about replacing developers with AI, but about creating a symbiotic relationship where human expertise and AI assistance combine to build better software faster than ever before.